Server CPU vs. Desktop CPU (Exact Difference?

[Supermansdead81]

I have tried to research this subject and keep coming up with differences between just a server system and a home desktop computer. I want to know what the differences are between the CPU's. I thought this may fit better in general hardware since I'm not specifing what brand because it is a general question. If you need an example then I mean what would be the advantage of using an Opteron or vice versa a Woodcrest chip over using an Athlon X2 or vice versa a Conroe based Processor. I've built 3 computers so far and have used a normal desktop CPU in everyone ranging from athlon t-birds to p4's and have never actually thought about what the difference might've been if I had used a server class processor. You know maybe I could get more performance going that route in a normal build but I can't find a clear answer. Please help and thanks to anyone that does.

 

[drizzt81]

The difference varies widely.

As of recently, a major differentiating factor is that desktop CPUs will not work in a multi-socket platform. There are rumors about extended reliability testing from the manufacturer.

 

[EQTakeOffense]

As of recently, a major differentiating factor is that desktop CPUs will not work in a multi-socket platform. There are rumors about extended reliability testing from the manufacturer.

That's the answer in a nutshell.

Usually high-end server chips get newer technology first such as Woodcrest's 1333Mhz FSB. The major downfall to this, is server hardware is usually very standard, lacking alot of features and/or overclocking ability but again, high-end server motherboards have newer technology as well such as FBDIMM. They also usually cost quite a bit more then their desktop counterparts.

Their main advantage is the ability to support multiple processors, which depending on your application can be good or a waste of money.

There are some Opterons and Xeons which are the same socket as their desktop equiliants. Usually there's not a whole lot of difference between these two, usually the biggest being price. The second is that sever class hardware goes through a more rigrious binning process and as such *MIGHT* overclock better.

 

[movax]

The difference has begun to widen now more than it used to be. (That's why people could turn Athlon XPs into Athlon MPs, etc.). And the old P2 & P3 Xeons were identical to their desktop counterparts except for Intel unlocking the necessary scalar parts of the Xeon dies.

Nowadays, its almost always a different socket (currently Socket 771 & Socket F), and the CPUs might differ in cache sizes & FSB as previously mentioned.

RAM types also differ, though mainly the difference is ECC/Registered, though Intel Woodcrests/Clovertowns are the only platforms that use FB-DIMMs atm. Higher-quality silicon/extended testing is also seemingly performed, which is why Opteron 150s/165s were overclocker favorites.

 

[zero_vertical]

server chips usually are tested longer, and usually require registered memory, which is more expensive.

 

[Dan_D]

server chips usually are tested longer, and usually require registered memory, which is more expensive.

Intel chips have never required registered memory. The motherboards that support Xeon CPUs do however require registered memory most of the time.